2.1.

Fabrication of Microaxicon Fiber Tip

Single mode optical fiber appropriate for 980 nm was used for preparing the axicon-tip fiber. One end of the mechanically cleaved bare single mode fiber was dipped into 50% hydrofluoric (HF) acid containing a protective layer (e.g., Toluene) at the top [Fig. 1a]. The tip formation takes place at the interface of the etching liquid and the protecting overlayer. As described by Hoffmann ¹⁶ the cone angle of the fiber tip is determined by the contact angle of HF with the fiber. The etching process is self-terminating and the cone angle is influenced by the liquid used as the top protection layer. Figure 1b shows a microscopic image of an axicon tip, prepared using a single stage chemical etching process. For fabrication of a large cone angle tip, the two-step etching technique¹⁷ was employed. After the first selective chemical etching step, the tapered region has a small cone angle. The second selective etching resulted in large cone angle axicon tip [Fig. 1c]. Figure 1d shows the SEM image of such a tip.

Fig. 1

(a) Time-lapse schematic of the chemical etching process. (b) Microscopic image of the axicon tip fabricated using single-stage self-terminating etching process. (c) Microaxicon tip fabricated using two-step etching process. (d) SEM micrograph of a microaxicon tip.

2.2.

Experimental Setup and Schematic of Fiber-Optic Tweezing and Two-Photon Excitation

The schematic experimental setup for fiber optic tweezers and two-photon excitation is shown in Fig. 2a. The output of the Ti:Sapphire laser (Maitai HP, Newport Spectra-Physics, Inc.) was expanded (BE) and coupled into the cleaved end of the axicon-tip single mode optical fiber using a 10× microscope objective (O). The microaxicon tip end of the fiber was held by an XYZ mechanical manipulator (or hand) and inserted into the cuvette (C). A side-viewing 10× microscope objective was used to image the microaxicon tip, trapped microspheres, and the two-photon fluorescence onto a video camera. In some cases, the two-photon fluorescence was observed by the naked eye by using a laser-safety eye glass or using a 100× microscope objective. The suspension of fluorescent polystyrene beads, dispersed in distilled water, was placed in the cuvette (C), and transilluminated using a white light source. An infrared cut-off filter was used to block the NIR laser beam reaching the CCD. The images of trapping and two photon excitation (TPE) of objects were digitized using a frame grabber and computer. The image analysis was done using Image J software.

Fig. 2

(a) The schematic experimental setup for fiber optic tweezers and two-photon excitation. L: Ti: Sapphire laser; BE: beam expander; S: Shutter; O: 10× microscope objective; C: cuvette; MO: microscope objective; CF: cut-off filter. Ray propagation for a microlens (b) versus microaxicon (c). (d) Schematic of two-photon fluorescence excitation in the trapped object along the optic axis by the Bessel–Gauss beam.

3.1.

Measurement of Beam Profile Through Microaxicon

An axicon is a specialized type of lens having a conical surface, which can be used to turn a Gaussian beam into a Bessel beam, with greatly reduced diffraction. Recent studies have shown¹⁵ that conical optics can focus better than spherical ones with the smallest optical confinement zones. Figures 2b and 2c show a schematic of ray propagation through a microlens versus microaxicon. No central ray along the optical axis, in the case of a microaxicon [Fig. 2c], ensures the least scattering force as compared to axial gradient force. Due to this special nature of the Bessel−Gauss beam (having a small high-intensity region along the Z-direction), relatively low NA (or less diverging beam) can achieve single beam optical tweezers as compared to traditional Gaussian beam optical tweezers. The two-photon excitation inside the trapped object occurs along the optic axis throughout the propagation distance [Fig. 2d], which depends on the tip cone angle of the microaxicon.

For different cone angle tips, the transverse beam profiles and propagation distance were measured by using a microscope objective, translated over a 20 μm range with a resolution of 1 μm. The measured beam profiles at different distances from the tip showed truncated Bessel-like beam profiles with few concentric rings. For a very small cone angle tip, the rays coming out from radially opposite locations of the tip do not intersect in contrast to the situation shown in Fig. 2c. Further, transmittance of the beam through the smaller microaxicon cone angle tip (<30 deg) fiber was found to decrease substantially, which can be attributed to an increase in total internal reflection at the fiber tip. Though these tips might be well-suited for near-field scanning microscope applications, two-photon excitation, and 3D-trapping requires higher transmittance and the highest intensity spot away from the tip. For a very large cone angle (>90 deg) tip, though transmittance is high, the propagation distance of the emanating laser beam profile is too large to generate sufficient axial gradient force for efficient 3D trapping. Therefore, for optimal two-photon fluorescence excitation and 3D trapping, anintermediate cone angle, i.e., between 40 to 60 deg, was used.

3.2.

Simulation of Beam Profile Through Microaxicon

Truncated-Bessel (or Bessel Gauss) beams, having different propagation characteristics, can be generated by engineering the axicon microstructure on the tip of the single mode fiber. For simulation of a beam profile emanating from the axicon microstructure, let us start with the fundamental mode of the fiber, [TeX:] $E_{{\rm fund}} \approx \exp ({^{ - (X^2 + Y^2)} /\omega ^2 })$ $E_{\mathrm{fund}}\approx \exp \left(-(X^2+Y^2)/{\omega }^2\right)$ where ω is the mode field size. The electric field propagating from the tip at any observation point (x ₀,y ₀) can be described¹⁸ as:

Fig. 3

(a) Nomenclature for the microaxicon. (b) Simulated beam profile of the laser beam (λ = 900 nm) transmitted through the axicon tip (T = 5.4 μm, r = 3 μm, θ = 29 deg) at distance of ∼10 μm. The scale in X and Y axis is in micrometers.

3.3.

Radiation Forces Due to Axicon Tip Fiber Tweezers in Rayleigh Regime

The optical trapping forces on the Rayleigh particles¹⁹ due to the beam emanating from the microaxicon fiber tip can be calculated using the transverse beam profiles [Fig. 3b] obtained at different axial distances. Considering a dielectric sphere (radius a = 50 nm, dielectric constant ε₁) in a medium (dielectric constant ε₂; magnetic permeability μ₂) being illuminated by a linearly polarized (parallel to the x axis) z-propagating beam from the microaxicon. The Rayleigh particle in the instantaneous electric field [TeX:] $\vec E(r,t)$ $\vec{E}(r,t)$ acts as a point dipole, whose dipole moment (in MKS units) is given by

The other component is a gradient force due to the Lorentz force acting on the dipole induced by the electromagnetic field. By using the electric dipole moment of Eq. 2 as an electrostatics analogue of the electromagnetic wave, the instantaneous gradient force is defined by

Fig. 4

Plot of radiation forces on a 50 nm polystyrene particle due to trapping beam power of 100 mW. Arrow points to stable axial trapping position.

Figure 4 shows the calculated scattering and axial gradient force as well as the total force along an axial direction on a polystyrene particle (n = 1.59) of radius 50 nm due to a laser beam power of 100 mW propagating through the axicon tip (T = 5.4 μm, r = 3 μm, θ = 29 deg). From Fig. 4, it is clear that the particle will be stably trapped along the axial direction (in addition to transverse directions by gradient force) at a distance of ∼ 6 μm (arrow marked) from the tip of the axicon. This is consistent with the fact that the beam width is minimal at a distance of ∼5 μm from the tip of the axicon [Fig. 3b] and the stable axial trapping position is shifted away from the axicon tip due to the scattering force. This stable axial trapping position and the magnitude of the axial trapping force were found to depend on the cone angle of the axicon as well as its refractive index and size of the particle. Therefore, the particle is expected to be trapped in this axial region. This stable trapping point and the magnitude of the trapping force were found to depend on the cone angle as well as refractive index of the axicon and size of the particle.

4.1.

Trapping and TPE of Fluorescent Objects Using Axicon Tip Single Optical Fiber

Theoretical simulation of forces exerted by the beam from the microaxicon tipped single fiber on dielectric objects showed that the axial gradient force can overcome the scattering force even for nanoobjects, thus enabling single beam three-dimensional trapping using fiber-optic axicon with 20 deg < θ < 40 deg. Though fs tweezing was as effective as cw tweezing, the Ti:Sapphire laser was operated in fs mode only for the time period for TPF recording, so as to cause least damage² to the trapped sample. Figure 5 shows two-photon fluorescence excitation [Fig. 5b] while trapping fluorescent polystyrene particle [Fig. 5c] in a single fiber fs optical tweezers was realized using an axicon tip with θ = 20 deg [Fig. 5a]. Quite stable two-photon excited fluorescence intensity was observed over the extended observation period (>5 min), under an inverted microscope. As discussed earlier, an intermediate axicon cone angle was found to provide better trapping stability. The two-photon fluorescence intensity map inside the particle was found to have a peak at an axial distance of 3.5 μm. However, the simulations showed that the beam waist is smallest at a distance of ∼3 μm from the tip of the axicon for θ = 20 deg. The small deviation from the simulated beam profile can be attributed to the presence of the particle acting as a microlens and refocusing the laser beam.

Fig. 5

(a) Bright field image of an axicon tip with θ = 20 deg. (b) Two-photon excitation in the trapped bead. (c) Bright field image also showing the two-photon fluorescence from the bead.

It may be noted that in contrast to two-photon excitation using a high NA microscope objective, a relatively longer streak of fluorescence was observed along the microsphere length [Fig. 5c] for microaxicon-based excitation. For θ = 20 deg, the propagation distance was theoretically estimated to be ∼4 μm, which matched the observed fluorescence. This has a distinct advantage for high signal-to-noise ratio detection of cells and molecules in the trapped region. However, this compromises the axial resolution. For higher axial resolution, smaller cone angle micro-axicon may be used. Indeed near-field trapping can be achieved¹⁴ using very small cone angle fiber axicons and therefore it can achieve two-photon excitation with high depth resolution.

4.2.

In-Depth Trapping and Two Photon Excitation of Trapped Fluorescent Microscopic Objects

Figure 6 shows trapping and two photon excitation of trapped fluorescent microspheres at depth exceeding few cm. The axicon tip [Fig. 6a] was immersed in the fluorescent microsphere suspension as shown in Fig. 6b. The side-viewing objective (10×, NA = 0.25) was used to image the tip and the trapped fluorescent particles inside the cuvette at a depth >1 cm [Fig. 6b]. Figure 6c shows the image of TPE fluorescence from a trapped 3 μm polystyrene particle at laser beam power of 84 mW in mode-locked condition. The highly-localized TPF was clearly visible to the naked eye. Figure 6d (Video 1) shows bright field image of trapping of the particle along with TPF using the side-viewing objective. With use of an IR cut-off filter, the TPF from the trapped particle could be filtered from the scattered light [Fig. 6e]. Figure 6f shows a zoomed view of the trapped and two-photon excited fluorescent particle. The bright field image of two particles trapped along the axial direction is shown in Fig. 6g. Figure 6h shows the TPF from both the trapped particles. The axial positioning of the two particles will be influenced by the propagation distance of the Bessel–Gauss beam and also by mutual interaction due to optical binding.¹⁴ Further, axial optical binding is known to depend on the size of the particle and therefore it will influence the distance between the two particles. Owing to the fact the these events occurred at a large depth from all sides of the cuvette, the side viewing objective with low NA limited us from resolving the distance between the two optically bound particles.

Video 1

Two-photon fluorescence from particles, trapped fiber optically inside a cuvette at a depth of centimeters. (QuickTime, 236 KB) v1"

Fig. 6

(a) Bright field images of the axicon tip inside cuvette. (b) The actual arrangement of axicon tip inside the cuvette at a depth >1 cm and the side-viewing objective. (c) TPE fluorescence of a trapped 3 μm polystyrene particle at laser beam power of 84 mW, as visible to the naked eye. (d) Bright field image of trapping along with TPF. (e) TPF from the trapped particle with IR cut-off filter. (f) Zoomed image of trapped and two-photon excited fluorescent particle with IR filter. (g) Bright field image of two trapped particles along axial direction with TPF. (h) TPF from the trapped particles with IR cut-off filter. (i) Zoomed image of trapped and two-photon excited fluorescent particles with IR filter. (j) TPF intensity as a function of average power of the fs laser beam and the fitted curve using a quadratic equation.

The fiber optic trapping was so stable that the TPF was observed by naked eye when holding the fiber in hand. Further, the trapped particle could be moved in 3D by slow maneuvering of the fiber. Figure 6i shows a zoomed view of the trapped and two-photon excited fluorescent particles with IR filter. The fact that a similar level of TPF intensity was observed from both the trapped particles indicates that the particle near the axicon tip might have refocused the laser beam onto the second particle [Fig. 6i]. Figure 6j shows variation of the TPF intensity from the trapped particle as a function of average power of the fs laser beam. Fitting the data to a quadratic equation showed a two-photon nature of the process.

It may be noted that a significantly large observation time could be achieved due to reduced photobleaching in the case of TPF excitation. The combination of the two noncontact methods in a single assay offers a powerful tool for studying cellular systems in large depths not possible earlier, by allowing simultaneous correlations made between molecular changes reported by fiber optic TPF, and micromechanical transitions probed by the fiber optical tweezers. The combined use of TPE and optical trapping using a single fiber optical device will enable in-depth analysis of nonadherent samples in microfluidic as well as bulk samples for in situ diagnostics. With the help of a 2×1 fiber optic splitter, the axicon-tip fiber, used for trapping, can collect the two-photon excited fluorescence (instead of the side-viewing imaging CCD used in the present study) for further analysis. Illustration of the concept for highly-sensitive detection of fluorescently-stained bacteria by two-photon excitation subsequent to concentration of the bacteria using fiber optic tweezers is shown in Fig. 7.

Fig. 7

Illustration of the concept for highly-sensitive detection of fluorescently-stained bacteria by two-photon excitation subsequent to concentration of the bacteria using fiber optic tweezers.